Skip to content

Latest commit

 

History

History
313 lines (181 loc) · 8.67 KB

INSTALL.rst

File metadata and controls

313 lines (181 loc) · 8.67 KB

Installation

Introduction

ANUGA is a python package with some C extensions (and an optional fortran extension). This version of ANUGA is run and tested using python 3.7 - 3.10

Dependencies

ANUGA requires python 3.X (X>6) and the following python packages:

numpy scipy matplotlib pytest cython netcdf4 dill future gdal \
pyproj pymetis Pmw mpi4py ipython meshpy Pmw pymetis utm

and

backport.zoneinfo 

if using python <= 3.8.

ANUGA is developed on Ubuntu and so we recommend Ubuntu as your production environment (though ANUGA can be installed on MacOS and Windows using Miniconda or MiniForge)

Install with MiniForge3 (Ubuntu)

A clean way to install the dependencies for ANUGA is to use Anaconda, or Miniconda Python distributions by Continuum Analytics.

Using a conda installation has the advantage of allowing you to create multiple python environments and is particularly useful if you want to keep multiple versions of ANUGA

Indeed the most stable install is via the conda-forge channel which is easily available using the Miniforge. In particular the installation of the gdal and mpi4py modules are more stable.

These conda environments do not require administrative rights to your computer and do not interfere with the Python installed in your system.

Install the latest version of Miniforge from https://github.com/conda-forge/miniforge or use, for instance, wget to download the latest version via:

wget -O Miniforge3.sh "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
bash Miniforge3.sh

If you don't have wget you can install it via:

sudo apt-get update -q
sudo apt-get install wget

Once Miniforge is installed, we can now create an environment to run ANUGA.

Create a conda environment anuga_env (or what ever name you like):

conda update conda
conda create -n anuga_env python=3.8 anuga mpi4py
conda activate anuga_env

Note we have also installed mpi4py to allow anuga to run in parallel. On some systems you may need to manually install mpi4py to match the version of mpi you are using.

This has setup a conda environment anuga_env using python 3.8. (ANUGA has be tested on 3.7, 3.8. 3.9.)

We are now ready to use ANUGA.

You can test your installation via:

conda activate anuga_env
python -c "import anuga; anuga.test()"

Install with MiniForge using ANUGA github repository (Ubuntu)

If you want to use the very latest version of ANUGA within a conda environment then we need to download the github version of ANUGA.

First install the latest version of Miniforge from https://github.com/conda-forge/miniforge or use, for instance, wget to download the latest version via:

wget -O Miniforge3.sh "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
bash Miniforge3.sh

and now create and activate a conda environment with ANUGA's dependencies

conda create -n anuga_env python=3.9 gxx pip wheel scipy numpy cython netcdf4 pytest nose matplotlib gdal dill future gitpython mpi4py utm Pmw pymetis meshpy 
conda activate anuga_env

Now we need to download the ANUGA source code from github

git clone https://github.com/anuga-community/anuga_core.git

and then install ANUGA

cd anuga_core
pip install -e .

This installs ANUGA to be "editable" so you should be able to develop code in the anuga_core directories (at your own risk).

Finally it is sensible to test the installation.

pytest --pyargs anuga

Updating

From time to time you might like to update your version of anuga to the latest version on github. You can do this by going to the anuga_core directory and pulling the latest version and then reinstalling via the following commands:

cd anuga_core
git pull
pip install -e .

And finally check the new installation by running the unit tests via:

pytest --pyargs anuga 

Installing on Ubuntu_20.04 and 22.04 using script install_ubuntu.sh

For Ubuntu 20.04 and 22.04 you can install ANUGA and its dependencies into a python virtual environment via a simple bash script.

First from your home directory run the following command which will download anuga to a directory `anuga_core`:

git clone https://github.com/anuga-community/anuga_core.git

Then the following will install dependencies, install anuga and run the unit tests:

bash anuga_core/tools/install_ubuntu.sh

Note: This script will only work for Ubuntu 20_04 and 22_04.

Note: Part of the bash shell will run as sudo so will ask for a password. If you don't like this, you can run the package installs manually, see the commands in the scripts anuga_core/tools/install_ubuntu_20._04.sh or anuga_core/tools/install_ubuntu_22._04.sh as appropriate.

This script also creates a python3 virtual environment anuga_env. You should activate this virtual environment when working with ANUGA, via the command:

source ~/anuga_core/anuga_env/bin/activate

You might like to add this command to your .bashrc file to automatically activate this python environment.

Updating

From time to time you might like to update your version of anuga to the latest version on github. You can do this by going to the anuga_core directory and pulling the latest version and then reinstalling via the following commands:

Activate the environment if necessary:

source ~/anuga_core/anuga_env/bin/activate

Then update ANUGA to latest version:

cd anuga_core
git pull
pip install -e .

And finally check the new installation by running the unit tests via:

pytest --pyargs anuga 

Windows 10 Install using 'Ubuntu on Windows'

Starting from Windows 10, it is possible to run an Ubuntu Bash console from Windows. This can greatly simplify the install for Windows users. You'll still need administrator access though. First install an ubuntu 20_04 subsystem. Then just use your preferred ubuntu install described above.

Windows Installation using MiniForge

We have installed ANUGA on windows using miniforge.

You can download MiniForge manually from the MiniForge site https://github.com/conda-forge/miniforge:

Alternatively you can download and install miniforge via CLI commands:

Run the following powershell instruction to download miniforge.

Start-FileDownload "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Windows-x86_64.exe" C:\Miniforge.exe; 

From a standard cmd prompt then install miniconda via:

C:\Miniconda.exe /S /D=C:\Py
C:\Py\Scripts\activate.bat

Install conda-forge packages:

conda create -n anuga_env python=3.8  anuga mpi4py
conda activate anuga_env

You can test your installation via:

python -c "import anuga; anuga.test()"

Installing GDAL on Ubuntu using apt and pip

ANUGA can be installed using the python provided by the Ubuntu system and using pip.

First set up a python virtual environment and activate via:

python3 -m venv anuga_env
source anuga_env/bin/activate

A complication arises when installing the gdal package. First install the gdal library, via:

sudo apt-get install -y gdal-bin libgdal-dev

We need to ascertain the version of gdal installed using the following command:

ogrinfo --version

THe version of gdal to install via pip should match the version of the library. For instance on Ubuntu 20.04 the previous command produces:

GDAL 3.0.4, released 2020/01/28

So in this case we install the gdal python package as follows

pip install gdal==3.0.4

Now we complete the installation of ANUGA simply by:

pip install anuga

If you obtain errors from pip regarding "not installing dependencies", it seems that that can be fixed by just running the pip install anuga again.